diff options
Diffstat (limited to 'fs')
63 files changed, 958 insertions, 500 deletions
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 3f75895c919..7301cdb4b2c 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -179,60 +179,74 @@ static int __add_prelim_ref(struct list_head *head, u64 root_id, static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path, struct ulist *parents, int level, - struct btrfs_key *key, u64 wanted_disk_byte, + struct btrfs_key *key_for_search, u64 time_seq, + u64 wanted_disk_byte, const u64 *extent_item_pos) { - int ret; - int slot = path->slots[level]; - struct extent_buffer *eb = path->nodes[level]; + int ret = 0; + int slot; + struct extent_buffer *eb; + struct btrfs_key key; struct btrfs_file_extent_item *fi; struct extent_inode_elem *eie = NULL; u64 disk_byte; - u64 wanted_objectid = key->objectid; -add_parent: - if (level == 0 && extent_item_pos) { - fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); - ret = check_extent_in_eb(key, eb, fi, *extent_item_pos, &eie); + if (level != 0) { + eb = path->nodes[level]; + ret = ulist_add(parents, eb->start, 0, GFP_NOFS); if (ret < 0) return ret; - } - ret = ulist_add(parents, eb->start, (unsigned long)eie, GFP_NOFS); - if (ret < 0) - return ret; - - if (level != 0) return 0; + } /* - * if the current leaf is full with EXTENT_DATA items, we must - * check the next one if that holds a reference as well. - * ref->count cannot be used to skip this check. - * repeat this until we don't find any additional EXTENT_DATA items. + * We normally enter this function with the path already pointing to + * the first item to check. But sometimes, we may enter it with + * slot==nritems. In that case, go to the next leaf before we continue. */ - while (1) { - eie = NULL; - ret = btrfs_next_leaf(root, path); - if (ret < 0) - return ret; - if (ret) - return 0; + if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) + ret = btrfs_next_old_leaf(root, path, time_seq); + while (!ret) { eb = path->nodes[0]; - for (slot = 0; slot < btrfs_header_nritems(eb); ++slot) { - btrfs_item_key_to_cpu(eb, key, slot); - if (key->objectid != wanted_objectid || - key->type != BTRFS_EXTENT_DATA_KEY) - return 0; - fi = btrfs_item_ptr(eb, slot, - struct btrfs_file_extent_item); - disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); - if (disk_byte == wanted_disk_byte) - goto add_parent; + slot = path->slots[0]; + + btrfs_item_key_to_cpu(eb, &key, slot); + + if (key.objectid != key_for_search->objectid || + key.type != BTRFS_EXTENT_DATA_KEY) + break; + + fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); + disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); + + if (disk_byte == wanted_disk_byte) { + eie = NULL; + if (extent_item_pos) { + ret = check_extent_in_eb(&key, eb, fi, + *extent_item_pos, + &eie); + if (ret < 0) + break; + } + if (!ret) { + ret = ulist_add(parents, eb->start, + (unsigned long)eie, GFP_NOFS); + if (ret < 0) + break; + if (!extent_item_pos) { + ret = btrfs_next_old_leaf(root, path, + time_seq); + continue; + } + } } + ret = btrfs_next_old_item(root, path, time_seq); } - return 0; + if (ret > 0) + ret = 0; + return ret; } /* @@ -249,7 +263,6 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, struct btrfs_path *path; struct btrfs_root *root; struct btrfs_key root_key; - struct btrfs_key key = {0}; struct extent_buffer *eb; int ret = 0; int root_level; @@ -294,19 +307,9 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, goto out; } - if (level == 0) { - if (ret == 1 && path->slots[0] >= btrfs_header_nritems(eb)) { - ret = btrfs_next_leaf(root, path); - if (ret) - goto out; - eb = path->nodes[0]; - } - - btrfs_item_key_to_cpu(eb, &key, path->slots[0]); - } - - ret = add_all_parents(root, path, parents, level, &key, - ref->wanted_disk_byte, extent_item_pos); + ret = add_all_parents(root, path, parents, level, &ref->key_for_search, + time_seq, ref->wanted_disk_byte, + extent_item_pos); out: btrfs_free_path(path); return ret; diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index e616f8872e6..12394a90d60 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -37,6 +37,7 @@ #define BTRFS_INODE_IN_DEFRAG 3 #define BTRFS_INODE_DELALLOC_META_RESERVED 4 #define BTRFS_INODE_HAS_ORPHAN_ITEM 5 +#define BTRFS_INODE_HAS_ASYNC_EXTENT 6 /* in memory btrfs inode */ struct btrfs_inode { diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index 9cebb1fd6a3..da6e9364a5e 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -93,6 +93,7 @@ #include "print-tree.h" #include "locking.h" #include "check-integrity.h" +#include "rcu-string.h" #define BTRFSIC_BLOCK_HASHTABLE_SIZE 0x10000 #define BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE 0x10000 @@ -843,13 +844,14 @@ static int btrfsic_process_superblock_dev_mirror( superblock_tmp->never_written = 0; superblock_tmp->mirror_num = 1 + superblock_mirror_num; if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE) - printk(KERN_INFO "New initial S-block (bdev %p, %s)" - " @%llu (%s/%llu/%d)\n", - superblock_bdev, device->name, - (unsigned long long)dev_bytenr, - dev_state->name, - (unsigned long long)dev_bytenr, - superblock_mirror_num); + printk_in_rcu(KERN_INFO "New initial S-block (bdev %p, %s)" + " @%llu (%s/%llu/%d)\n", + superblock_bdev, + rcu_str_deref(device->name), + (unsigned long long)dev_bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + superblock_mirror_num); list_add(&superblock_tmp->all_blocks_node, &state->all_blocks_list); btrfsic_block_hashtable_add(superblock_tmp, diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index d7a96cfdc50..15cbc2bf4ff 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -467,6 +467,15 @@ static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info, return 0; } +/* + * This allocates memory and gets a tree modification sequence number when + * needed. + * + * Returns 0 when no sequence number is needed, < 0 on error. + * Returns 1 when a sequence number was added. In this case, + * fs_info->tree_mod_seq_lock was acquired and must be released by the caller + * after inserting into the rb tree. + */ static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags, struct tree_mod_elem **tm_ret) { @@ -491,11 +500,11 @@ static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags, */ kfree(tm); seq = 0; + spin_unlock(&fs_info->tree_mod_seq_lock); } else { __get_tree_mod_seq(fs_info, &tm->elem); seq = tm->elem.seq; } - spin_unlock(&fs_info->tree_mod_seq_lock); return seq; } @@ -521,7 +530,9 @@ tree_mod_log_insert_key_mask(struct btrfs_fs_info *fs_info, tm->slot = slot; tm->generation = btrfs_node_ptr_generation(eb, slot); - return __tree_mod_log_insert(fs_info, tm); + ret = __tree_mod_log_insert(fs_info, tm); + spin_unlock(&fs_info->tree_mod_seq_lock); + return ret; } static noinline int @@ -559,7 +570,9 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info, tm->move.nr_items = nr_items; tm->op = MOD_LOG_MOVE_KEYS; - return __tree_mod_log_insert(fs_info, tm); + ret = __tree_mod_log_insert(fs_info, tm); + spin_unlock(&fs_info->tree_mod_seq_lock); + return ret; } static noinline int @@ -580,7 +593,9 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info, tm->generation = btrfs_header_generation(old_root); tm->op = MOD_LOG_ROOT_REPLACE; - return __tree_mod_log_insert(fs_info, tm); + ret = __tree_mod_log_insert(fs_info, tm); + spin_unlock(&fs_info->tree_mod_seq_lock); + return ret; } static struct tree_mod_elem * @@ -1023,6 +1038,10 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info, looped = 1; } + /* if there's no old root to return, return what we found instead */ + if (!found) + found = tm; + return found; } @@ -1143,22 +1162,36 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, return eb_rewin; } +/* + * get_old_root() rewinds the state of @root's root node to the given @time_seq + * value. If there are no changes, the current root->root_node is returned. If + * anything changed in between, there's a fresh buffer allocated on which the + * rewind operations are done. In any case, the returned buffer is read locked. + * Returns NULL on error (with no locks held). + */ static inline struct extent_buffer * get_old_root(struct btrfs_root *root, u64 time_seq) { struct tree_mod_elem *tm; struct extent_buffer *eb; - struct tree_mod_root *old_root; - u64 old_generation; + struct tree_mod_root *old_root = NULL; + u64 old_generation = 0; + u64 logical; + eb = btrfs_read_lock_root_node(root); tm = __tree_mod_log_oldest_root(root->fs_info, root, time_seq); if (!tm) return root->node; - old_root = &tm->old_root; - old_generation = tm->generation; + if (tm->op == MOD_LOG_ROOT_REPLACE) { + old_root = &tm->old_root; + old_generation = tm->generation; + logical = old_root->logical; + } else { + logical = root->node->start; + } - tm = tree_mod_log_search(root->fs_info, old_root->logical, time_seq); + tm = tree_mod_log_search(root->fs_info, logical, time_seq); /* * there was an item in the log when __tree_mod_log_oldest_root * returned. this one must not go away, because the time_seq passed to @@ -1166,22 +1199,25 @@ get_old_root(struct btrfs_root *root, u64 time_seq) */ BUG_ON(!tm); - if (old_root->logical == root->node->start) { - /* there are logged operations for the current root */ - eb = btrfs_clone_extent_buffer(root->node); - } else { - /* there's a root replace operation for the current root */ + if (old_root) eb = alloc_dummy_extent_buffer(tm->index << PAGE_CACHE_SHIFT, root->nodesize); + else + eb = btrfs_clone_extent_buffer(root->node); + btrfs_tree_read_unlock(root->node); + free_extent_buffer(root->node); + if (!eb) + return NULL; + btrfs_tree_read_lock(eb); + if (old_root) { btrfs_set_header_bytenr(eb, eb->start); btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV); btrfs_set_header_owner(eb, root->root_key.objectid); + btrfs_set_header_level(eb, old_root->level); + btrfs_set_header_generation(eb, old_generation); } - if (!eb) - return NULL; - btrfs_set_header_level(eb, old_root->level); - btrfs_set_header_generation(eb, old_generation); __tree_mod_log_rewind(eb, time_seq, tm); + extent_buffer_get(eb); return eb; } @@ -1650,8 +1686,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, BTRFS_NODEPTRS_PER_BLOCK(root) / 4) return 0; - btrfs_header_nritems(mid); - left = read_node_slot(root, parent, pslot - 1); if (left) { btrfs_tree_lock(left); @@ -1681,7 +1715,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, wret = push_node_left(trans, root, left, mid, 1); if (wret < 0) ret = wret; - btrfs_header_nritems(mid); } /* @@ -2615,9 +2648,7 @@ int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key, again: b = get_old_root(root, time_seq); - extent_buffer_get(b); level = btrfs_header_level(b); - btrfs_tree_read_lock(b); p->locks[level] = BTRFS_READ_LOCK; while (b) { @@ -5001,6 +5032,12 @@ next: */ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) { + return btrfs_next_old_leaf(root, path, 0); +} + +int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, + u64 time_seq) +{ int slot; int level; struct extent_buffer *c; @@ -5025,7 +5062,10 @@ again: path->keep_locks = 1; path->leave_spinning = 1; - ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (time_seq) + ret = btrfs_search_old_slot(root, &key, path, time_seq); + else + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); path->keep_locks = 0; if (ret < 0) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 0236d03c673..fa5c45b3907 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2753,13 +2753,20 @@ static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans, } int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path); -static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p) +int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, + u64 time_seq); +static inline int btrfs_next_old_item(struct btrfs_root *root, + struct btrfs_path *p, u64 time_seq) { ++p->slots[0]; if (p->slots[0] >= btrfs_header_nritems(p->nodes[0])) - return btrfs_next_leaf(root, p); + return btrfs_next_old_leaf(root, p, time_seq); return 0; } +static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p) +{ + return btrfs_next_old_item(root, p, 0); +} int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path); int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf); int __must_check btrfs_drop_snapshot(struct btrfs_root *root, diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index c18d0442ae6..2399f408691 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -1879,3 +1879,21 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) } } } + +void btrfs_destroy_delayed_inodes(struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + struct btrfs_delayed_node *curr_node, *prev_node; + + delayed_root = btrfs_get_delayed_root(root); + + curr_node = btrfs_first_delayed_node(delayed_root); + while (curr_node) { + __btrfs_kill_delayed_node(curr_node); + + prev_node = curr_node; + curr_node = btrfs_next_delayed_node(curr_node); + btrfs_release_delayed_node(prev_node); + } +} + diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 7083d08b2a2..f5aa4023d3e 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -124,6 +124,9 @@ int btrfs_fill_inode(struct inode *inode, u32 *rdev); /* Used for drop dead root */ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); +/* Used for clean the transaction */ +void btrfs_destroy_delayed_inodes(struct btrfs_root *root); + /* Used for readdir() */ void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, struct list_head *del_list); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 7ae51decf6d..7b845ff4af9 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -44,6 +44,7 @@ #include "free-space-cache.h" #include "inode-map.h" #include "check-integrity.h" +#include "rcu-string.h" static struct extent_io_ops btree_extent_io_ops; static void end_workqueue_fn(struct btrfs_work *work); @@ -2118,7 +2119,7 @@ int open_ctree(struct super_block *sb, features = btrfs_super_incompat_flags(disk_super); features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF; - if (tree_root->fs_info->compress_type & BTRFS_COMPRESS_LZO) + if (tree_root->fs_info->compress_type == BTRFS_COMPRESS_LZO) features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO; /* @@ -2575,8 +2576,9 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) struct btrfs_device *device = (struct btrfs_device *) bh->b_private; - printk_ratelimited(KERN_WARNING "lost page write due to " - "I/O error on %s\n", device->name); + printk_ratelimited_in_rcu(KERN_WARNING "lost page write due to " + "I/O error on %s\n", + rcu_str_deref(device->name)); /* note, we dont' set_buffer_write_io_error because we have * our own ways of dealing with the IO errors */ @@ -2749,8 +2751,8 @@ static int write_dev_flush(struct btrfs_device *device, int wait) wait_for_completion(&device->flush_wait); if (bio_flagged(bio, BIO_EOPNOTSUPP)) { - printk("btrfs: disabling barriers on dev %s\n", - device->name); + printk_in_rcu("btrfs: disabling barriers on dev %s\n", + rcu_str_deref(device->name)); device->nobarriers = 1; } if (!bio_flagged(bio, BIO_UPTODATE)) { @@ -3400,7 +3402,6 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, delayed_refs = &trans->delayed_refs; -again: spin_lock(&delayed_refs->lock); if (delayed_refs->num_entries == 0) { spin_unlock(&delayed_refs->lock); @@ -3408,31 +3409,37 @@ again: return ret; } - node = rb_first(&delayed_refs->root); - while (node) { + while ((node = rb_first(&delayed_refs->root)) != NULL) { ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); - node = rb_next(node); - - ref->in_tree = 0; - rb_erase(&ref->rb_node, &delayed_refs->root); - delayed_refs->num_entries--; atomic_set(&ref->refs, 1); if (btrfs_delayed_ref_is_head(ref)) { struct btrfs_delayed_ref_head *head; head = btrfs_delayed_node_to_head(ref); - spin_unlock(&delayed_refs->lock); - mutex_lock(&head->mutex); + if (!mutex_trylock(&head->mutex)) { + atomic_inc(&ref->refs); + spin_unlock(&delayed_refs->lock); + + /* Need to wait for the delayed ref to run */ + mutex_lock(&head->mutex); + mutex_unlock(&head->mutex); + btrfs_put_delayed_ref(ref); + + spin_lock(&delayed_refs->lock); + continue; + } + kfree(head->extent_op); delayed_refs->num_heads--; if (list_empty(&head->cluster)) delayed_refs->num_heads_ready--; list_del_init(&head->cluster); - mutex_unlock(&head->mutex); - btrfs_put_delayed_ref(ref); - goto again; } + |